Zed — High-Performance AI Code Editor

What it is

Zed is a high-performance code editor built from scratch in Rust by the team that created Atom and Tree-sitter. It uses GPU acceleration for rendering, resulting in frame rates and input latency that traditional Electron-based editors cannot match. The editor includes built-in multiplayer editing and an integrated AI assistant.

Zed targets developers who find VS Code or JetBrains IDEs too slow and want a native-speed editor with modern AI features. It runs on macOS and Linux.

How it saves time or tokens

Zed's AI assistant connects directly to language models (OpenAI, Anthropic, or custom providers) within the editor. Instead of switching to a separate chat window, you highlight code, ask the assistant to refactor or explain it, and the result appears inline. The editor's speed means less waiting for UI responsiveness, and the multiplayer feature eliminates the back-and-forth of screen sharing during pair programming.

How to use

1. Install Zed on your platform:

# macOS
brew install --cask zed

# Linux
curl -f https://zed.dev/install.sh | sh

1. Configure your AI provider in Settings (Cmd+,):

{
  "assistant": {
    "default_model": {
      "provider": "anthropic",
      "model": "claude-sonnet-4-20250514"
    }
  }
}

1. Use the AI assistant: select code, press Ctrl+Enter to ask a question, or open the assistant panel for longer conversations.

Example

// Zed settings.json — configure AI and editor preferences
{
  "assistant": {
    "default_model": {
      "provider": "openai",
      "model": "gpt-4o"
    }
  },
  "theme": "One Dark",
  "buffer_font_size": 14,
  "format_on_save": "on",
  "vim_mode": true
}

Related on TokRepo

• AI coding tools -- Compare AI-powered editors and coding assistants
• Featured workflows -- Discover curated developer tools and configurations

Common pitfalls

• Zed's extension ecosystem is smaller than VS Code's. Check if your must-have language servers and extensions are available before switching full-time.
• Multiplayer editing requires a Zed account and internet connection. For offline-only workflows, the core editor works fine but collaboration features are unavailable.
• AI assistant configuration varies by provider. Anthropic models need an API key set in the settings, and token costs apply based on your provider's pricing.